Chemical vapour deposition of germanium-containing films by IR laser-induced decomposition of ethoxy(trimethyl)germane.код для вставкиСкачать
APPI-IED ORGANOMETALLIC CHEMISTRY, VOL. 9, 667-673 (1995) Chemical Vapour Deposition of Germanium-containing Films by IR Laser-induced Decomposition of Ethoxy(trimethy1)germane Radek Fajgar,* Zdengk Bastl,t Jaroslav TIaskalS and Josef Pola*§ Academy of Sciences of the Czech Republic, * Institute of Chemical Process Fundamentals, 165 02 Prague 6-Suchdol, Rozvojova 135, t J. Heyrovsky Institute of Physical Chemistry, 182 23 Prague 8, and 3 Institute of Inorganic Chemistry, 25 068 Re?, near Prague, Czech Republic Carbon Dioxide (CO, ) laser-induced decomposition of ethoxy(trimethy1)germane(ETG) results in a substantial stripping of organic substituents from germanium and leads to deposition of organogermanium films, the composition of which is dependent on the mode of laser irradiation. Direct absorption of laser radiation in ETG affords material rich in Germanium, while a sulfur hexafluoride (SF,)-photosensitized process produces a deposit composed of Germanium, Carbon, Hydrogen and Oxygen. The deposited materials can be modified by chemical reactions with acetic anhydride and atmospheric moisture. Keywords: chemical vapour deposition; laserinduced decomposition; organogermaniurn film; ethoxy(trirnethy1)germane INTRODUCTION The preparation of polymeric organogermanium oxides, analogous to silicones, has long been of interest in organogermanium chemistry, but traditional methods of preparing these organogermanium polymers using hydrolysis of dihalogermanes,1-3 dialkoxygermanes,4-6 organogermanium trichloridesG9 or tetraalkoxygermanes" have been recognized to yield only low-molecular-weight, water-soluble oligomers, germanium dioxide or digermoxanes. An alternative approach to the production of polymeric organo-oxogermanes is the radical decomposition of alkyl(a1koxy)germanes in the gas phase. Previous studies on metallo-organic chemical vapour deposition (MOCVD) reveal § Author to whom correspondence should be addressed. CCC 0268-2605/95/080667-07 01995 by John Wiley & Sons, Ltd. that pyrolytic" and plasma-induced" decomposition of tetraethoxygermane affords only germanium dioxide, but laser-induced decomposition of tetramethoxygermane yields films of reactive organo-oxogermanium polymers. l3 These different reflect the commonly shared view that the nature of deposited materials is affected by the structure of the gaseous precursor and by the conditions of precursor decomposition; this has been well documented even for laser-induced MOCVD using organogermanes. 13-" In a continuation of our previous studies on IR laser-induced MOCVD of polyorganooxogermanes from tetrameth~xygermane'~ and of germanium from tetramethylgermane," we report in this paper the gas-phase carbon dioxide (CO,) laser-induced decomposition of ethoxy( trimethy1)germane (ETG) and assess the use of direct infrared multiphoton decomposition (IRMPD) as well as sulphur hexafluoride (SF6) photosensitized decomposition (PSD) of ETG as a technique for preparation (deposition) of polymeric organo-oxogermanes. EXPERIMENTAL Laser irradiation experiments were performed on gaseous samples of ETG and ETG-SF, compounds contained in a cylindrical glass reactor (10 cm x 3.6 cm i.d.), equipped with sodium chloride (NaCl) windows, a PTFE stopcock and a sleeve with rubber septum. A grating-tuned transversely excited atmospheric (TEA) C 0 2 laser" was operated at a repetition frequency of 1 Hz (energy in pulse 0.28 J cm-') on the R(12) line of the 00"1-02"0 transition (1073.2 cm-') to achieve absorption in ETG. A grating-tuned continuousReceived 21 September 1994 Accepted 27 February I995 668 wave COz lase?" (output 10 W) operating on the transition P(20) line of the 00'1-02"O (944.2 cm-') was chosen for the irradiation of the ETG-SF, mixture when absorption in SF,, was effective. The laser beam of the pulsed and continuous radiation was focused at the centre of the horizontally positioned reactor using a Ge or NaCl lens, and the substrate (NaCI, potassium bromide (KBr) glass or aluminium) for the deposit was housed 1.5 cm beneath the focal point. The laser beam energy was measured with a laser energy pyroceramic sensor (Charles University, ml-1JU model) or a Coherent Model 201 power meter, and the laser line used for the irradiation was verified with a model 16-A spectrum analyser (Optical Engineering Co.). The samples for laser irradiation were prepared by a standard vacuum-line technique and the pressure of ETG was measured by a Barocel pressure transducer (model 1570). The IR spectra before and after irradiation were recorded with a Specord 75 model (Zeiss) IR spectrometer. The depletion of ETG was monitored at 810 and 1040 cm-'. Gaseous products of the laser-induced decomposition of ETG were identified by their absorption spectra (methane 1300 cm-', ethene 950 cm-', ethyne 730 cm-' and acetaldehyde 1746cm-'), by their mass fragmentation and by their retention times. For the latter purpose, helium was expanded into the reactor to atmospheric pressure and gaseous samples were injected into a G U M S (Shimadzu Q P 1000) quadrupole mass spectrometer (column 1.2 m long packed with Porapak P, programmed temperature 25-160 "C). The amounts of gaseous compounds were determined by using the absorptivity of the diagnostic (strong, non-overlapping) bands and the comparison with molar absorptivities measured with authentic samples (cm-I, lO'XkPa-'cm-'): ETG, 1040, 64 and 810, 65; CH,, 1300, 30; C2HZ,730, 94: C2H4, 950, 70; CH,CHO, 1746,58. The deposit on glass or NaCl substrates was investigated by means of X-ray photoelectron spectroscopy, scanning electron microscopy, UV/Vis and IR spectroscopy. ESCA measurements were made using a VG ESCA 3 MkII electron spectrometer. The pressure of residual gases during accumulation of the spectra was in the 10-6Pa range. The measurements were performed using AIK, (1486.6 eV) radiation. The spectrometer was operated in the fixed-analyser transmission mode with a pass energy of 20 eV giving a resolution of 1.1eV on R. FAJGAR, Z . BASTL, J . TLASKAL AND J . POLA the Au f7,2 line. The preparation chamber of the spectrometer was equipped with a cold cathode ion gun. The spectra of Ge 2p, 3d, i" 1s and 0 1s photoelectrons and Ge L3M45M,5Auger electrons were measured. The ratios of atomic concentrations were determined by correcting the photoelectron peak areas for their cro$s-sections*' and by taking into account the dependence of the photoelectron mean free path and analyser transmission on electronic kinetic The overlapping spectral features we re resolved into individual components of GausGan-Lorentzian shape using a modified version of the damped non-linear squares procedure published by Hughes and Sexton.24For binding energy data we estimated the error limit of +0.2eV. The estimated accuracy of the calculated ratios of atomic concentrations amounted to k 10'%. UV/Vis absorption spectra of the deposit were measured, using a Hewlett-E'ackard 8451A spectrometer, in the range 120-900 nm. Scanning electron microscopy (ISEM) studies of the deposit were performed O I I an ultra-high vacuum Tesla BS 350 instrument equipped with an energy-dispersive analyser of !<-ray radiation, Edax 9100/65. An ECON detectcx in the shield mode (plastic window) was used for qualitative determination of light elements. The morphology of the samples was investigated mostly using an accelerating voltage of 4 kV. ETG samples as well as authentic samples of trimethylgermane and tetramethj,,lgermane were prepared as reported'" and distilled under vacuum before use. Sulphur hexafluoride was a commercial sample from Fluka. '' RESULTS AND DISCUSSION Infrared multiphoton decomposition (IRMPD) Focused irradiation by the C 0 2 laser (0.28 J in pulse) in the strong absorption band ( v ~mode) ~ - ~ of ETG at 1073 cm-' results (Figs 1 and 2) in the depletion of ETG and formation of methane, ethene, ethyne, acetaldehyde, carbon monoxide and a compound whose structure was tentatively assigned to (H2CH3Ge)?0[infrared absorption at 2035 cm-'; mass spectrum - CH4(CH3)2Ge20+ (characteristic mass range 190-202)]. A significant amount of a solid brown material, deposited CVD OF GERMANIUM-CONTAINING FILMS 669 II u1 0.4. 500 0 iooo Number of pulses Figure3 Distribution of volatile products in IRMPD of ETG: B,ETG; A , C2H4; 0, CZHZ; 0, CH3CHO; 0, CHI). -d I 1000 2000 3000 Wavenumber, cm-1 Figure 1 IR spectra of the irradiated (900 pulses, (a)] and initial (b) ETG, and of the solid deposit (c) and the deposit treated with acetic anhydride (d). all over the inside of the reactor, is produced concomitantly. Mass balance measurement (IR spectral determination of amounts of gaseous products) (Fig. 3) is consistent with approximately 80% of carbon being used for the formation of the gaseous products. A very weak absorption band at 2035 cm-' and the high molar absorptivity of organogermanium hydrides at this ~ a v e l e n g t h lead ' ~ us to believe that the volatile digermoxane is formed only at pressures lower than 0.01 kPa, which implies that almost all the germanium, less than 20% of the carbon and less than 60% of the oxygen from the ETG is incorporated in the deposited material. This estimation is consistent with the stoichionietry of the deposit as determined by X-ray photoelectron spectroscopy (XPS) analysis (Table 1). Photosensitizeddecomposition (PSD) Focused irradiation by the continuous-wave (cw) CO, laser on ETG-SF, (each component 0.7 kPa) mixtures (incident energy 40 W cm-') tuned to Table 1 XPS core level binding energies, Auger parameters (eV) and composition of the deposits. Source of deposit IRMPD IRMPD~ PSD PSDb IRMPD/Ac,O Auger Ge 2pyz Ge 3d parameter' 1219.3 1217.9 1220.3 1218.1 1218.2 1220.3 IRMPDIAcZOd 1217.3 1219.3 1221.0 1217.2 1220.4 Retention time Figure2 Typical GC-MS trace of the irradiated ETG (a) and ETG-SF, (b). Peak identification: 1, air; CO; 2, CHI; 3, C2H4,C,H,; 4, CH3CHO; 5, ETG, C,H,OH; 6, (CH,),Ge, (CH,),GeH, ETG, C,H,OH; 7, (CH,H2Ge),0. - 1174.3 1174.7 1171.4 1174.5 -' Gel Gel oC,300 Gel oC4,OI Gel ,C1 8008 Gel o c 0 0 0 1 o -' Gel o C U 4 0 1 o 1174.7 1170.9 Clean Gee GeO?' - 29.6 33.3 Value of Auger parameter based on Ge 3d peak. After 10 min of sputtering with argon ions (energy 5 keV, ion current approx. 40 PA). 'Because of overlap of Auger peaks which could not be separated, the Auger parameter values were not calculated. After 1min of sputtering with argon ions (energy 3.5 keV, ion current 40 PA). Reference (commercial) samples of polycrystalline germanium and GeO, respectively. a 1 30.3 29.9 32.4 30.1 29.9 32.3 Stoichiometry R FAJGAR, Z. BASTL. J TLASKAL AND J . POLA 670 the strong absorption band of SF, at 944 cm-' has a similar effect to direct absorption of the laser radiation in ETG. SF, acts as an energyconveying agent" '' and induces homogeneous decomposition of ETG. Irradiation times shorter than 40 s were sufficient to achieve approximately 60% decomposition. Gaseous products formed included those observed upon IRMPD of ETG and also tetramethylgermane and trimethylgermane (Fig. 2b). Formation of a brown powder-like material was observed as well, but its amount was smaller than in IRMPD. The similarity of the main volatile products in IRMPD and particle-size determination (PSD) of ETG indicates that the major steps contributing to the decomposition are the same. The analysis of the deposited materials by XPS (Table l), as well as the amounts of gaseous products (Fig. 3) are in line with a substantial cleavage of organic moieties from germanium. The cleavage of the Ge-0 and Ge-C bonds is easier than that of the 0 - C bond (homolytic bond dissociation energy 300 kJ mol-'," 240 kJ mol-',") and 340 kJ mol-', respectively). The initial Ge-C and Ge-0 cleavages in ETG are obviously followed by radical reactions with parent ETG as hydrogen abstraction from the OCH,CH, (major, Eqn ) or OCH,CH, (minor, Eqn [ 3 ] )units, loss of acetaldehyde (major, Eqn ), or ethene (minor Eqn [ S ] ) . These products can also be formed by direct four-centre /?-elimination reactions (Eqns , ). Second-order radical reactions, e.g. disprois less likportionation of an ethoxy ely. Decomposition of acetaldehyde3' (Eqn 181) and ethanol' (Eqn ) can occur, too. The proposed steps are given in Eqns [ 1]-. * I I Ge--OCH,CH, + CH; + 'OCH,CH3 CH,' CH; I t E T G - + CH,-Ge' + (CH,),GeOCH,CH,-+CH, + (CH ,),GeOCH'CH, + (CH,),GeOCH,CH,-.CH, + (CH,) ,GeOCH2CH,' I [11 PI (31 + CH,CHO  (CH,),GeOCH2CH2'-t (CH,),GeO' + C2H4  (CH,),GeOCH'CH,-+ (CH,),Ge' (CH3)3GeOCH2CHI+(CH3),GeH+ CH,CHO [61 (CH7)7GeOCH,CH3-+ (CH,)Ge( )H + C,H, CH,CHO+ CH, + CO  PI CH,CH,OH+ C2H4 + H,O [91 Formation of trimethylgermarre might also occur via molecular expulsion of ethene oxide which rearranges" 3 3 into acetaldehyde Eqn [lo]), rather than by an ~nlikely~"'~ hydrogen abstraction by the (CH,),Gc' radical. 'H- C H ~ CH~CHO Properties of the deposit ESCA, FTIR and SEM analyses of the deposits afforded by IRMPD and PSD of ETG reveal that the materials contain germanium, carbon, hydrogen and oxygen and that they have different properties depending on the way they were produced. Energy dispersive X-ray spectrometry (EDX)-SEM analysis of the bulk material (of ca 0.3 pm thickness) is in line with greater amounts of the deposited agglomerates and with a higher proportion of germanium in the deposit from IRMPD compared with that from PSD (Fig. 4). These results conform with the stoichiometry of superficial (up to 5 nm) layers analysed by ESCA (Table 1).The deposit from PSD contains neither fluorine nor sulphur, which shows no chemical involvement of the sensitizer. The analysis of the layers beneath those removed by ion sputtering reveals that the material deposited by IRMPD is very similar to elemental germanium, while that obtained by PSD contains, apart from germanium, carbon and oxygen also. The amounts of oxygen in superficial layers of both deposits is higher than it would be if it corresponded to the observed Ge-0 cleavage/formation of acetaldehyde and ethanol, indicating that superficial layers incorporate oxygen from the atmosphere. The Auger parameters for the unchanged and ion-sputtered deposits show that the IRMPD-originated material contains elemental germanium, while that obtained from PSD incor- 67 1 CVD OF GERMANIUM-CONTAINING FILMS b a Figure 4 SEM image and EDX-SEM trace of the deposit afforded by PSD (a) and IRMPD (b). porates superficial GeO, , elemental germanium being incorporated only in deeper layers. Optical absorption data of the film obtained by IRMPD (Fig. 5 ) show that the absorption edge (transparency) is not reached at wavelengths lower than 800 nm. The deposits do not possess good adhesion to glass, sodium chloride or aluminium. They are soluble in acetone and tetrahydrofuran to form brown solutions, but are insoluble in ethanol and alkanes. The deposits obtained from IRMPD and PSD exert a similar pattern of infrared absorption bands (770, 850, 1230, 1370-1400, 2900 and .-0 + . 3 2970 cm-'; Fig. lc); this is c ~ n s i s t e n t ~ 'with . ~ * the occurrence of Ge-0-C and C-H bonds. In order to prove that the deposits contain Ge-OC2Hs groups, we examined the reactivity of the layers with water vapour and acetic anhydride. It is known that water reacts" with alkoxygermanes to yield GeOz and ethanol, and that acetic anhydride reacts (R. Fajgar, unpublished results) with ETG giving ethyl acetate and acetoxy(trimethy1)germane. We found that the introduction of air into the evacuated reactor containing a deposit from IRMPD leads to formation of ethanol (new bands at 1055, 1230 and 1400cm-' and a broad band centred at 3400cm-'), and that treatment of the same deposit with acetic anhydride vapour (exposure to acetic anhydride, evacuation and exposure to air) alters the infrared spectrum of the deposit due to the occurrence of a stronger band at 1400cm-' and a new band at 1600 cm-l (Fig. Id) which can U P, L 0 I 2 < +Ge-OCn2CH3- I - - J A d - I 15 D - Figure 5 Absorption spectrum of the deposit from IRMPD. I +Ge-OH+ GH~OH I H20 +Ge-OC(O)CH3+Ge-OH+ c: 2 0 Wavelength, nm H20 -CH3COOC2Hj I Scheme 1 1 I CH,CO*H R. FAJGAR, Z . BASTL. J. TLASKAL AND J . POLA 672 I I I I I I 1210 1220 1230 BINDING ENERGY (eV) (a) 280 285 290 BINDING ENERGY (eV) (b) - l " " I ' " ' I ' " ' be assigned to adsorbed acetic acid. These changes appear to be due to the reactions shown in Scheme 1 and they reveal that the deposited materials can be chemically modified. The incorporation of oxygen into the deposit from IRMPD has also been confirmed by XPS analysis (Table 1); the equal amounts of germanium and oxygen in beth superficisl and deeper layers of the treated deposit show that the incor'poration of oxygen can be enhanced by chemical treatment, and that the materials perhaps have a porous structure facilitating the penetration of reactive vapours. The approximately ten-fold increase in the oxygen content in deeper layers of the deposit from IRMPD experimerits upon treatment with acetic anhydride indicates that the incorporation of oxygen cannot be solely due to reactions of Ge-OC2H, groups, m d that it is perhaps caused by reactions of reactive (naked) germanium centres. The nature of tnese reactions as well as that of the centers are at present unknown. The shape of the germantum core level spectra measured after sample treatment with acetic anhydride suggests they are composed of two peaks (Fig. 6a; Table 1) corresponding to the order of increasing binding energy likely to be ascribed to Ge-C and Ge-0 and/or Ge-OH bonds, respectively. After ion sputtering, the additional peak corresponding to elemental germanium appears in the G e 2p3,* photoelectron spectrum. This assignment is consistent with C 1s and 0 1s spectra (Fig. 6b and c, respectively). Acknowledgement The authors thank the Grant Agency of the Academy of Sciences of the Czech Republic for financial support (grant no. 47266). I REFERENCES 525 530 535 540 BINDING ENERGY (eV) (C) Figure6 Spectra of Ge 2p,, (a), C 1s (b), and 0 1s (c) photoelectrons of the deposit treated with acetic anhydride before ( 1 ) and after (2) argon ion sputtering. 1. E. G . Rochow, in Comprehensive Inorganic Chemistry, Vol. 2, J . C. Bailar, H. J . EmelCus, R. Nyholm, A. F. Trotman-Dickenson, eds. Pergamon Press, Oxford, 1973, pp. 1-41. 2. M. P. Brown and E. G. Rochow, 1. A m . Chem. SOC. 82, 4166 (1960). 3. W. Metlesies and H. Zeiss, J . A m . Cherq. SOC. 82, 3324 (1960). 4. F. Rijkens, Orgunogermanium Compour:ds, Germanium Research Committee, Utrecht, 1960. 5. R. Schwarz and K. G . Knauff, Z . Anopg. Allg. Chem. 275, 193 (1954). 6. L. Ross and M. Drager, 2.Naturforsch., Teil B 39, 868 (1984). CVD O F GERMANIUM-CONTAINING FILMS 7. K. Haberle and M. Drager, 2. Naturforsch., Teil B 39, 1541 (1984). 8. H. Puff, K. Brdun, S. Franken, T. R. Kok and W. Schuh, J. Organomet. Chem. 349, 293 (1988). 9. H. Puff, K. Braun, S. Franken, T. R. Kok and W. Schuh, J. Organomet. Chem. 335, 167 (1987). 10. M. Lesbre, P. Mazerolles and J. Satge, The Organic Compounds of Germanium, Wiley, London (1971). 1 1 . V. G. Minkina, Izv. Akad. Nauk SSSR, Neorg. Muter. 24, 803 (1988), Chem. Abstr. 109, 46855e. 12. M. Ya Ivanov, T. N. Kupryashkina, G. Aloyan, N. A. Ovsyannikov, E. A. Ryabenko and A. Efremov, Sint. Soed. Plasma Soderzh. Uglevodorody, 141 (1985); Chem. Abstr. 97, 1183542. 13. J. Pola, R. Fajgar, Z. Bastl and L. Diaz, J. Muter. Chem. 2 , 961 (1992). 14. J. Pola, J. P. Parsons and R. Taylor, J. Chem. Soc., Faraday Trans. 88, 1637 (1992). 15. J. Pola and R. Taylor, J . Organomet. Chem. 437, 271 (1992). 16. J. Pola, J. P. Parsons and R. Taylor, J. Muter. Chem. 2, 1289 (1992). 17. M. Jakoubkovi, Z. Bastl, P. Fiedler and J. Pola, Infrared Phys. Technol. 35, 633 (1994). 18. A. E. Stanley, R. A. Johnson, J. B. Turner and A. H. Roberts, Appl. Spectrosc. 40,374 (1986). 19. J. Pola and S. Simeonov, J. Chem. SOC.,Perkin Trans. 2 , 101 (1991). 20. J. Pola, M. FarkaEovi, P. Kubit and A. Trka, J . Chem. Soc., Faraday Trans. 1 8 0 , 1499 (1984). 21. J. H. Scofield, J. Electron Spectrosc. Relat. Phenom. 8, 129 (1976). 22. M. P. Seah and W. A. Dench, Surface Interface Anal. 1,1 (1979). 673 23. L. T. Weng, G. Vereecke, M. J . Genet, P. Bertrand and W. E. E. Stone, Surface Interface Anal. 20, 179 (1992). 24. A. E. Hughes and B. A. Sexton, J. Electron Spectrosc. Relat. Phenom. 46, 31 (1988). 25. W. M. Shaub and S. H. Bauer, Int. J . Chem. Kinet. 7,509 ( 1975). 26. D. K. Russell, Chem. SOC.Rev. 19, 407 (1990). 27. Yu. Kh. Shaulov, A. K. Fedorov, G. Ya. Zueva, G. V. Borisyuk and V. G. Gendel, Zh. Fiz. Khim. 99, 2081 (1970). 28. S. W. Benson, Thermochemical Kinetics, Wiley, New York, 1976. 29. P. Gray, R. Shaw and J. C. Thynne, Prog. React. Kinet. 4, 90 (1967). 30. M. J . Gibian and R. C. Corky, Chem. Reu. 73, 441 (1973). 31. T. Berces in Comprehensive Chemical Kinetics, Bamford, C . H. and Tipper, C. F. H. (eds), Vol. 5, Elsevier, Amsterdam, 1972. 32. W. H. Richardson and H. E. O’Neal, in Comprehensiue Chemical Kinetics, C. H. Bamford and C. F. H. Tipper (eds), Vol. 5, Elsevier, Amsterdam, 1972. 33. M. L. Neufeld and A. T. Blades, Can. J. Chem. 41,2956 (1963). 34. J . Lucztyk, B. Maillard, S. Deycard, D. A . Lindsay and K. U. Ingold, I . A m . Chem. SOC. 105, 3578 (1983). 35. J. W. Wilt, J. Lusztyk, M. Peeran and K. U. Ingold, J. Am. Chem. Soc. 110, 281 (1988). 36. D. C. Carson, K. U. Ingold and L. C. Bray, Int. 1. Chem. Kinet. 1, 315 (1969). 37. F. Glockling, The Chemistry of Germanium, Academic Press, London, 1969. 38. V. C. Farmer and J. D. Russel, Spectrochim. Acta 18,461 (1962).